On Tue Sep 29 2009 at 06:51:58 +0000, David Holland wrote: > On Mon, Sep 28, 2009 at 11:38:49PM +0300, Antti Kantee wrote: > > > Avoid nasal demons. Code of the form > > > > > > vput(vp); > > > error = VFS_VGET(vp->v_mount, ...); > > > > > > just isn't right. Because of vnode caching this *probably* never bit > > > anyone, except maybe under very heavy load, but still. > > > > Unbitten more likely because checkpath is only for rename and rename has > > approximately 57 billion extra references to the vnode when calling it. > > "but still". > > Well, not there; that code is/was executed on each step ascending the > path up from the destination directory. That's fairly likely to hit > some otherwise idle places.
Oh, right. But I don't think this looks like a possible explanation for the segfault in 42093 (which is most likely a reference counting fail)... but still, good catch ;)